Oligo(cyclohexylidene)s and oligo(cyclohexyl) as bridges for photoinduced intramolecular charge separation and recombination

Article abstract of DOI:10.1021/jp0271716

A series of semirigid donor-bridge-acceptor (D-B-A) molecules was synthesized to study the effect of the position and number of nonconjugated olefinic bonds in the bridge on the photoinduced charge-separation and charge-recombination kinetics. The molecules consist of a phenylpiperidine electron donor, an oligo-(cyclohexylidene) or oligo(cyclohexyl) bridge, and a dicyanovinyl acceptor. Partly saturated ter(cyclohexylidene) bridges were used as well. The edge-to-edge donor-acceptor separation of the compounds under study varies between 2.89 and 15.4 A. The replacement of a C-C single bond by an olefinic bond increases the rate of charge separation with a factor of 3.0 ± 0.8 per replaced bond. For all D-B-A compounds the extended fully charge-separated state folds to a compact charge-transfer (CCT) conformer. The rate of charge recombination (CR) of the CCT state increases with solvent polarity for those compounds having an olefinic bond located three σ bonds from the acceptor. Thus, while in cyclohexane the CR rate is equal for all compounds, in benzene the CR rate in compounds with an olefinic bond near the acceptor is 10 times larger than in compounds with a single bond instead. It is believed that a (virtual) charge-transfer state involving the radical cation of the olefinic bond and the radical anion of the acceptor (D-B^;+-A^-) is responsible for the enhanced CR process.

Full text of DOI:10.1021/jp0271716

3612
J. Phys. Chem. A 2003, 107, 3612-3624
Oligo(Cyclohexylidene)s and Oligo(Cyclohexyl)s as Bridges for Photoinduced Intramolecular
Charge Separation and Recombination†
Wibren D. Oosterbaan,‡ Carola Koper,‡ Thijs W. Braam,‡ Frans J. Hoogesteger,‡
Jacob J. Piet,‡,| Bart A. J. Jansen,‡ Cornelis A. van Walree,*,‡ H. Johan van Ramesdonk,§
Marijn Goes,§ Jan W. Verhoeven,§ Wouter Schuddeboom,| John M. Warman,| and
Leonardus W. Jenneskens*,‡
Debye Institute, Department of Physical Organic Chemistry, Utrecht UniVersity, Padualaan 8,
3584 CH Utrecht, The Netherlands, Laboratory of Organic Chemistry, UniVersity of Amsterdam,
Nieuwe Achtergracht 129, 1018 WS Amsterdam, The Netherlands, and Radiation Chemistry Department,
IRI, Delft UniVersity of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands
ReceiVed: October 9, 2002; In Final Form: February 14, 2003
A series of semirigid donor-bridge-acceptor (D-B-A) molecules was synthesized to study the effect of
the position and number of nonconjugated olefinic bonds in the bridge on the photoinduced charge-separation
and charge-recombination kinetics. The molecules consist of a phenylpiperidine electron donor, an oligo-
(cyclohexylidene) or oligo(cyclohexyl) bridge, and a dicyanovinyl acceptor. Partly saturated ter(cyclohexylidene)
bridges were used as well. The edge-to-edge donor-acceptor separation of the compounds under study varies
between 2.89 and 15.4 Å. The replacement of a C-C single bond by an olefinic bond increases the rate of
charge separation with a factor of 3.0 ( 0.8 per replaced bond. For all D-B-A compounds the extended
fully charge-separated state folds to a compact charge-transfer (CCT) conformer. The rate of charge
recombination (CR) of the CCT state increases with solvent polarity for those compounds having an olefinic
bond located three σ bonds from the acceptor. Thus, while in cyclohexane the CR rate is equal for all
compounds, in benzene the CR rate in compounds with an olefinic bond near the acceptor is 10 times larger
than in compounds with a single bond instead. It is believed that a (virtual) charge-transfer state involving
the radical cation of the olefinic bond and the radical anion of the acceptor (D-B•+-A•-) is responsible for
the enhanced CR process.
Introduction
groups, such as a dihalocyclopropane unit,7 which provide low
energy virtual states D•+-B•--A or D-B•+-A•-, can enhance
the rate of charge separation. Although a fast charge-separation
step is beneficial for the efficiency of photochemical energy
conversion, it should be realized that charge-recombination rates
can be enhanced by super-exchange involving these states as
well. For an extensive literature survey on the effect of
intermediate and virtual states on charge-separation and
-recombination rates the reader is referred to ref 8. π-Electron
containing systems in a bridge may also participate in a so-
called sequential mechanism9,10 in which the D•+-B•--A or
D-B•+-A•- states exist as real intermediates.
An interesting bridge is represented by the oligo(cyclohexy-
lidene) skeleton. Oligo(cyclohexylidene)s11 (Chart 1A) consist
of an array of cyclohexane rings which are linked at the 1 and/
or 4 positions by double bonds. The σ-π-topology in these
molecules is well capable of relaying through-bond interactions
between functionalities attached to the termini as has been
demonstrated12,13 by photoelectron spectroscopy in combination
with ab initio calculations. Self-assembled monolayers of R,ω-
sulfur-functionalized oligo(cyclohexylidene)s have been applied
as medium for electron transfer between CdSe quantum dots
and a gold electrode.14,15 The distance dependence of the rate
of ET of these systems was found to be comparable to or in
some cases even smaller than that16,17 of fully π-conjugated
bridges. In donor-acceptor-functionalized bi(cyclohexylidene)
D[3π3]A, as well as in bi(cyclohexyl) D[7]A (Chart 2),
photoinduced ET to a D•+-B-A•- state has been observed.18
Photoinduced long-range electron transfer (ET) has long been
recognized as the key process in the conversion of solar energy
to chemical energy as taking place in natural and artificial
photosynthetic systems. The study of various donor-bridge-
acceptor (D-B-A) molecules with rigid saturated hydrocarbon
bridges has contributed significantly to the understanding of the
electron-transfer process. The influence of factors as donor-
acceptor separation, reaction medium, and free energy on the
ET process has been studied extensively.1-4 In the case of
saturated hydrocarbon bridges, ET proceeds via a superexchange
or through-bond mechanism5 and its rate is related to the
magnitude of the coupling between the locally excited donor
or acceptor state and the charge-transfer (CT) state.
The observation that a bacteriochlorophyll (BChl) is located
between the initially excited bacteriochlorophyll dimer (P) and
the bacteriopheophytin (H) electron acceptor in the purple
bacterial photosynthetic reaction center, combined with the fact
that extremely fast ET from P to H occurs with a time constant6
of 3 ps over a 17 Å center-to-center distance, has stimulated
the research on the influence of intermediate π-systems on the
rate of ET. It has been found that π-systems as well as other
† “Oligo(cyclohexylidene)s” and “oligo(cyclohexyl)s” are shorthand
notations for compounds named oligo(cyclohexan-1,4-diylidene)s and oligo-
(cyclohexan-1,4-diyl)s, respectively, according to IUPAC nomenclature.
‡ Debye Institute.
§ Laboratory of Organic Chemistry.
| Radiation Chemistry Department.
10.1021/jp0271716 CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/04/2003

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3613
CHART 1: Oligo(cyclohexylidene)s (A) and
Oligo(cyclohexyl)s (B) with the r- and ω-Positions
Indicated
SCHEME 1: Synthetic Scheme for the Syntheses of
D[3π3π3]A, D[3π3π3π3]A, D[7π3]A, D[11]A, and
D[3π7]Aa
CHART 2: Structure of Donor-Bridge-Acceptor
Compounds and Model Systems Studied
The donor D in these compounds is of the N,N-dialkylaniline
type, while the acceptor A is a dicyanovinyl group. The part
between brackets denotes the bond topology of the bridge.
Although we were not able to determine the rate of charge
separation, it was found that in benzene the rate of charge
recombination in D[3π3]A is substantially (20 times) faster than
that in D[7]A.19 For D[3π3]A and D[7]A in the nonpolar solvent
cyclohexane evidence was also obtained for the occurrence of
large conformational changes (folding) accompanying charge
separation in the nonpolar solvent cyclohexane. While for D-
[3π3]A it could not be determined whether folding precedes or
follows charge separation, in the case of D[7]A a harpooning
mechanism,20-23 in which charge separation precedes the folding
process, was observed.
In the present study we report and discuss the photophysical
properties of a series of longer D-B-A compounds D[3π3π3]-
A, D[7π3]A, D[3π7]A, D[11]A, and D[3π3π3π3]A (Chart 2).
In these compounds we have systematically varied the number
and position of the olefinic bonds in the bridge as to be able to
address the influence of these parameters on the charge-
separation and -recombination processes. Furthermore, the
photophysical properties of D[3π3]A and D[7]A are studied in
more detail. In particular, the effect of solvent polarity on both
the charge-recombination process and the folding process is
discussed. The compounds were studied by means of steady-
state spectroscopic methods and time-resolved absorption and
fluorescence spectroscopy and microwave conductivity.
a Reagents: (a) (i-Pr)2NLi, THF; (b) Me2NCH(OCH2CMe3)2, MeCN;
(c) HCl, H2O, THF; (d) H2C(CN)2, NH4Ac, HOAc, benzene; (e) H2,
Pd/C, THF (EtOH); (f) 1. NaOH, H2O 2. HCl, H2O.
Results and Discussion
Synthesis. The syntheses of donor-bridge-acceptor com-
pounds D[3π3π3]A, D[3π7]A, D[7π3]A, D[11]A, and D-
[3π3π3π3]A are shown in Scheme 1. A key intermediate in
the synthetic routes is carboxylic acid 2e. Initially, it was
prepared by a reaction sequence involving catalytic hydrogena-
tion of ethyl 4-hydroxybenzoate to form ethyl 4-hydroxycyclo-
hexanecarboxylate.11,24 Since the yield of this step was irrepro-
ducible (0-95%), an alternative and more reliable route was
used. This route (Scheme 2) involves the reaction of diethyl
malonate with two equivalents of ethyl acrylate by 1,4-additions
forming tetraethyl pentane-1,3,3,5-tetracarboxylate (2a), which
was ring-closed by means of a Dieckmann condensation to
form25,26 triethyl 4-oxocyclohexane-1,1,3-tricarboxylate (2b).
This product was converted into 4-oxocyclohexanecarboxylic
acid (2c) by means of acidic hydrolysis and decarboxylation.
Subsequently, 2c was converted into 2e via 2d.

3614 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
SCHEME 2: Synthetic Scheme for the Synthesis of 2c
and 2ea
Figure 1. Anti-syn-equilibrium of a bi(cyclohexylidene).
TABLE 1: Center-to-Center Rc and Edge-to-Edge Re
Donor-Acceptor Distances from MM2-Optimized
Geometries of the D-B-A Compounds under Studya
compound
Rc (Å)
Re (Å)
D[3]A
D[3π3]A
4.16
8.3
[8.41]
7.5
2.89
7.02
[7.04]
6.3
a
a
s
D[7]A
8.6
7.34
[8.70]
12.5
11.7
11.5
11.6
12.6
10.9
12.3
11.2
13.1
16.7
15.4
[7.33]
11.19
10.52
10.45
10.36
11.39
9.98
11.37
10.05
11.84
15.37
14.24
D[3π3π3]A
aa
as
sa
ss
a
s
a
D[3π7]A
D[7π3]A
a Reagents: (a) EtONa, THF; (b) EtONa, diethyl ether; (c) HCl, H2O;
(d) 2,2-dimethyl-1,3-propanediol, H+, toluene; (e) 1. NaOH, MeOH 2.
HCl, H2O.
s
D[11]A
D[3π3π3π3]A
Alkylation of the dianion of carboxylic acid 2e with ketone
1 (Scheme 1) afforded â-hydroxy acid 3 which was decarboxy-
lated and dehydrated to yield acetal 4. Hydrolysis of 4 gave
ketone 5, which was condensated with malononitrile to afford
D[3π3π3]A. The same reaction sequence starting from ketones
5 and 9 furnished D-B-A compounds D[3π3π3π3]A and D-
[7π3]A, respectively. Catalytic hydrogenation of acetal 11
yielded 13 as a mixture of cis and trans isomers. Hydrolysis of
this mixture followed by isomer separation gave the trans isomer
of ketone 14, which upon Knoevenagel condensation afforded
D[11]A.
For the synthesis of D[3π7]A, synthon 18 was prepared as
shown in Scheme 1. First, 16a was prepared by coupling of the
dianion of 2e with 15, which was prepared by esterification of
4-oxocyclohexanecarboxylic acid (2c, Scheme 2) with neopentyl
alcohol. Dehydrative decarboxylation of the coupling product
then yielded 16a together with the byproducts 16b and 16c (see
Supporting Information). Catalytic hydrogenation of 16a gave
the cis and trans isomers of 17 in a 1:1 ratio (according to GC
analysis). Subsequent basic saponification followed by careful
acidification gave 18 as a mixture of cis and trans isomers. For
the synthesis of D[3π7]A, 18 had to be coupled to 19. In this
reaction it was found that the conditions under which dianion
formation of carboxylic acid 2e is readily achieved did not lead
to dianion formation of 18. This is probably caused by the low
solubility of the dilithium salt of the dianion of 18 in THF.
This solubility problem was solved by Trost and Tamaru27 for
a similar carboxylic acid by the use of THF-HMPA mixtures.
Although we found that the addition of HMPA indeed promoted
dianion formation of 18, in our hands the best results were
obtained by the use of TMEDA instead of HMPA. Dehydrative
decarboxylation of 20, followed by deprotection and condensa-
tion with malononitrile, finally gave D[3π7]A.
aaa
sss
a Distances between brackets are crystal data.35 The labels “a” and
“s” denote anti and syn geometry, respectively, with the first label
referring to the double bond closest to the donor.
are expected to be considerably more flexible. For the bridges
which contain an olefinic bond syn-anti isomerism (Figure 1)
is important. For 4,4′-di-tert-butylbi(cyclohexylidene) under
isomerization conditions an equilibrium composition with a anti/
syn ratio of 0.83/0.17 has been reported33 which corresponds
to a Gibbs free energy difference of 0.95 kcal mol-1. It is
reasonable to expect that also for the compounds in the present
study both syn and anti conformers are present in solution.
Support for the presence of different conformers in solution
of the compounds with bridges that contain an olefinic bond
was found18 in 1H NMR spectra of D[3π3]A, D[3π3], and [3π3]-
A. For these compounds a single, time-averaged chemical shift
is observed for the equatorial and axial protons attached to the
same carbon atom in the saturated hydrocarbon rings, whereas
for D[7]A, D[7], and [7]A distinct signals are observed for the
equatorial and axial protons attached to the same carbon atom.
This implies that compounds which contain an olefinic bond
are conformationally flexible on the NMR time scale, while the
conformation of their saturated analogues is frozen. Moreover,
it was found that in D[7]A and D[7] the phenyl group occupies
an equatorial position.18 Nevertheless, B3LYP calculations on
D[3π3] have shown34 that a conformer with the phenyl ring in
an axial position is sufficiently low in energy (0.8 kcal mol-1
above the energy minimum) to be accessible at room temper-
ature. The donor-acceptor distance in this conformer is,
however, only slightly different from that of the lowest energy
conformer.
All D-B-A compounds possess high melting points and low
solubilities in common organic solvents which decrease with
increasing molecular size.
Ground-State Conformational Properties and Donor-
Acceptor Distances. An important factor in the study of charge
separation in D-B-A compounds is the distance between donor
and acceptor, which should be well-defined. Therefore, rigid
bridges, such as norbornyl28 (see however ref 29), cubyl,30 and
steroid skeletons,31,32 have been applied. The oligo(cyclohexy-
lidene) and oligo(cyclohexyl) bridges used in the present study
As a consequence of the occurrence of syn-anti isomerism
(see Figure 1), the interchromophoric distance in the compounds
with double bonds in the bridge is not fixed. To get an estimate
of the variation in the interchromophoric distance molecular
mechanics (MM2) calculations have been performed on all
possible syn and anti conformers of the D-B-A compounds
with bridges that contain one or more olefinic bonds. In Table
1 the donor-acceptor edge-to-edge distances Re of the various
D-B-A compounds and their syn and anti conformers (if any)
are listed. The distance Re is measured as the distance between

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3615
the anilino nitrogen atom and the olefinic C atom of the acceptor
which is incorporated in the ring. As can be seen syn-anti
interconversion produces a variation in the edge-to-edge distance
of up to 1.4 Å (for D[3π7]A). In Table 1 also the center-to-
center distance Rc has been indicated. This is the distance
between the midpoint of the phenyl-C-N bond and the midpoint
of the central olefinic bond of the acceptor. For D[3π3]A and
D[7]A the calculated Re and Rc values can be compared with
their crystal structures.35 In the crystal, the (partly) saturated
six-membered rings of both molecules adopt a chair-type
conformation and the phenyl group adopts a (pseudo) equatorial
orientation. The bridge of D[3π3]A is in an anti conformation.
As can be seen in Table 1 the MM2 distances are in good
agreement with the crystal data.
Electronic Absorption Spectra. The absorption spectra of
D[3π3]A, D[7]A, D[3π3], D[7], [7]A, and [3π3]A have been
discussed before18,34 and will only be briefly reviewed here.
The N,N-dialkylanilino donor exhibits three absorption bands
in the region 195-350 nm. For D[7] in cyclohexane, these bands
are positioned at 203, 255 and 289 nm.18 The olefinic bond in
1,1′-bi(cyclohexylidene) has a maximum at 206 nm,36,37 while
the π*rπ-transition of the dicyanovinyl acceptor in [7]A gives
rise to a single maximum at 235 nm.18 In [3π3]A, besides these
two π*rπ-transitions, another band with a maximum at 274
nm in cyclohexane18 is present, which has been ascribed to a
transition involving charge transfer from the olefinic bond to
the dicyanovinyl acceptor.18,38
Figure 2. Absorption spectra of D[3π3π3]A (solid), D[7π3]A (dash),
[3π3π3]A (dash-dot), and D[3π3] (dot) in dichloromethane at 20 °C.
TABLE 2: Absorption Maxima λmax (nm) and Molar
Absorption Coefficients E (103 M-1 cm-1) Determined in
CH2Cl2 at 20 °C (range: 230-500 nm)
compound
λmax
ꢀ
D[3π3]
258.0
258.5a
248.0
245.5a
286.5
290.5
236.0
13.3
12.0
20.7
21.1
3.58
3.47
D[7]
D[3π3]A
D[7]A
[3π3]A
[3π3π3]A
[3π7]A
D[3π3π3]A
D[7π3]A
D[3π7]A
D[11]A
15.8
22.65b
20.1
The absorption spectrum of D[7]A can be described as the
sum of the spectra of D[7] and [7]A. Hence, no noticeable
ground-state interaction between the donor and acceptor in D-
[7]A is present. The same was concluded18 for D[3π3]A, but a
careful reexamination of the absorption and excitation spectra
of this compound revealed a very weak absorption tail extending
to about 355 nm in cyclohexane (estimated band maximum
between 300 and 320 nm). This absorption is neither present in
[3π3]A nor in D[3π3] and is therefore ascribed to a transition
involving charge separation from the phenylpiperidine donor
to the dicyanovinyl acceptor: D•+[3π3]A•- r D[3π3]A. It thus
appears that the presence of the olefinic bond in D[3π3]A
induces a ground-state interaction between the anilino donor
and the dicyanovinyl acceptor. For D[3]A, which has an
identical donor-acceptor pair, the position of the CT absorption
has been reported to be located at 342 nm in n-hexane,39,40 and
it has a considerably larger absorption coefficient. Due to the
larger distance between the charges, the CT state D•+[3π3]A•-
will be less stabilized than that of D[3]A, and hence a blue-
shift for the absorption maximum is expected,2 which is indeed
observed.
236.0
244.5a
245.5a
23.8
22.2
a Shoulder in the region 285-295 nm. b Not any maximum present,
ꢀ at 245 nm.
about one-third of the absorbed excitation light will produce a
locally excited donor and about two-thirds of the excitation light
will produce a “local” [3π•+3]A•- charge-transfer state. Starting
with these two locally excited states, in principle two processes
are possible: electron transfer, in which the electron residing
on the excited donor is transferred to the acceptor, and hole
transfer, in which the hole initially produced on an olefinic bond
is transferred to the N,N-dialkylanilino donor. In D[3π3]A
excitation in the 300-315 nm region will also directly produce
the fully charge-separated state.
Solvent Dependence of the Steady-State Fluorescence of
D[3π3]A and D[7]A. Fluorescence maxima νf and fluorescence
quantum yields Φf of all D-B-A compounds as well as those
of [3π3]A and [3π3π3]A are given in Table 3. The steady-
state fluorescence spectra of D[3π3]A and D[7]A, recorded in
solvents of different polarity, are presented in Figure 3. The
local donor emission, located at 336 nm for both D[3π3] and
D[7] in cyclohexane,18 is largely quenched for D[3π3]A
(g99.8%) and D[7]A (g97%) in all solvents.41 A comparison
between the emission spectra of D[3π3]A and those of [3π3]A
(not shown) reveals that emission from the partly charge-
separated state D[3π•+3]A•-, which is formed upon photoex-
citation of D[3π3]A as well (see above), is not present in any
solvent. This indicates that, in addition to electron transfer from
the excited phenylpiperidine donor to the dicyanovinyl acceptor,
also hole transfer from the olefinic bond to the phenylpiperidino
donor takes place. The fluorescence bands which remain are
ascribed to the fully charge-separated states D•+[3π3]A•- and
D•+[7]A•- for D[3π3]A and D[7]A, respectively.18
The absorption spectra of D[7π3]A, D[3π3π3]A, [3π3π3]-
A, and D[3π3] in dichloromethane are presented in Figure 2.
Absorption maxima and molar absorption coefficients are
tabulated in Table 2. The absorption spectrum of [3π3π3]A
shows an absorption maximum which is slightly red-shifted as
compared to that of [3π3]A. Furthermore, in [3π7]A CT
absorption bands could not be detected. The absorption spectra
of D[3π3π3]A, D[7π3]A, D[3π7]A, and D[11]A can be
reasonably well described as the sum spectra of separate donor
and acceptor compounds.
It is important to note that the UV spectra show that in the
wavelength range 300-315 nm, in which most samples have
been excited for fluorescence measurements, in the case of D-
[7]A, D[3π7]A, and D[11]A only the anilino donor is excited,
while in the case of D[3π3]A, D[3π3π3]A, and D[7π3]A only

3616 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
TABLE 3: Experimental Steady-State Charge-Transfer Fluorescence Maxima νf (in nm) (Quantum Yield Φf/10-3) of the
Compounds Given in Various Solvents at 20 °C Together with the Results of Solvatochromic Fits
solvent
cyclohexane
∆f
D[3π3]A
D[7]A
[3π3]A
D[3π3π3]A D[7π3]A
D[3π7]A
496 (27)
D[11]A
[3π3π3]A D[3π3π3π3]A
0.101 388 (3.0) 497a (28) 361 (1.2)
521 (32)
0.117 525 (3.0) 515 (23)
0.173 525a(5.5) 483 (17)
0.193 520a(4.7) 484 (17)
0.234 522 (3.8) 521 (20)
0.254 542 (2.9) 545 (12)
0.293 630
0.308 630 (0.3) 650
0.319
0.376
511 (48)
512 (18)
508 (4.7)
360 (2.0)
b
benzenec
416 (15)
378 (4.1)
381 (4.8)
389 (7.9)
409 (10)
468 (8.1)
464 (12)
461 (19)
512 (1.1)
574 (3.2)
550 (8.1)
554 (6.2)
558 (3.9)
576 (1.9)
f
576 (2.6)d
550 (4.4)d
555 (3.5)d
566 (2.1)d
578 (1.1)d
∼468 (1.4)g
548 (3.9)
531 (4.5)
534 (4.1)
537 (2.6)
555 (1.9)
f
547 (1.8)
415 (22)
∼415, 577d
di-n-pentyl ether
di-n-butyl ether
diisopropyl ether
diethyl ether
ethyl acetate
THF
∼500a (1.6) 379 (6.9)
b
518a,d (∼1.7) 382 (7.2)
b
540e (0.5)
552 (1.2)
f
389 (9.2)
410 (14)
468 (9.9)
465(13)
461 (25)
512 (1.0)
32.0 ( 0.9
b
b
650
∼468 (b)
∼635a
∼464 (1.3)g ∼650a (e0.1) f
∼465 (2.2)
CH2Cl2
butyronitrile
νf(0) (103 cm-1
)
)
30.8 ( 0.5 28.9 ( 1.3 32.0 ( 1.0 20.7 ( 0.4 20.8 ( 0.3 22.5 ( 1.0
-49.3 ( 2 -43.9 ( 5 -32.5 ( 3.6 -13.0 ( 2.2 -13.8 ( 1.5 -12.5 ( 2.3 -9.6 ( 1.4 -32.6 ( 3.5
0.997 0.974 0.959 0.958 0.982 0.954 0.989 0.961
20.7 ( 0.3
slope (103 cm-1
correlation coefficient
a Not used in the solvatochromic fit (see text). b Solubility too low. c Values obtained in benzene were not used in the solvatochromic comparison,
since it is known that the polarity of benzene is not well described by its ∆f. d Maximum obtained by deconvolution. e Value obtained by streak
camera measurement. f No CT fluorescence observed. g Residual acceptor emission.
Figure 3. Fluorescence spectra of D[3π3]A (A) and D[7]A (B) in various solvents (excited between 270 and 300 nm). The intensity of all spectra
is corrected for the absorbance (excited at 300 nm).
In cyclohexane, D[3π3]A clearly shows two CT bands
positioned at 388 and 520 nm. The 520 nm band has previously
been assigned to a folded or compact charge-transfer (CCT)
conformer of the D•+[3π3]A•- state.18 The remaining band at
388 nm (3.20 eV) is assigned to the extended CT state (ECT
state) of D•+[3π3]A•-.
of donor and acceptor, respectively, as determined in acetonitrile
(ꢀs ) 37). When it is assumed that λi′ is independent of Rc,
which is reasonable since the main portion of the structural
reorganization is expected to be located on the donor and
acceptor sites and not on the bridge, eq 2 predicts a blue shift
of the ECT emission maximum upon increasing Rc. For D[3]A
in cyclohexane an emission maximum of 450 nm40 (2.76 eV)
has been reported and hence the ECT state D•+[3π3]A•- of D-
[3π3]A is indeed destabilized with respect to that of D[3]A by
an amount of 0.44 eV.
The energy of the CT fluorescence is given by2
hνf ) ∆G - λi′ - λs + E00
(1)
where h is Planck’s constant, ∆G is the (negative) energy
difference between the relaxed locally excited state and the
relaxed CT state, λi′ is the (positive) inner reorganization energy
upon going from the Franck-Condon ground state formed upon
CT fluorescence to the relaxed ground state, while λs is the
solvent reorganization energy associated with this process. E00
is the zero-zero transition energy of the locally excited state
from which charge transfer takes place. In a model where the
donor and acceptor are represented by two spheres with mean
radius r separated by a center-to-center distance Rc and
immersed in a medium with refractive index n and permittivity
ꢀs eq 1 becomes2
While for D[3π3]A in cyclohexane two maxima are present,
one corresponding to an ECT species and one to a CCT species,
in the more polar solvents di-n-pentyl ether and di-n-butyl ether
only a single very broad fluorescence band is present (Figure
3A). The full widths at half-height ∆ν1/2 of the spectra in di-
n-pentyl ether and di-n-butyl ether are ∼7.4 and ∼7.2 × 103
cm-1, respectively, while values for CT fluorescence bands
typically range from 4.5 to 6.5 × 103 cm-1 42
. Furthermore, the
location of the band maximum seems to be virtually independent
of solvent polarity. These observations indicate that also in di-
n-pentyl ether and di-n-butyl ether emission is observed from
both ECT and CCT species. A similar broad band is present in
benzene which shows that even in this quasi polar solvent CCT
emission is observed. This supports our earlier presumption18
that the small excited-state dipole moment obtained with time-
resolved microwave conductivity (TRMC) in benzene is due
to the presence of folded conformers. Note that the occurrence
of two species in this solvent was confirmed with time-resolved
hνf ) F[Eox(D) - Ered(A)] - e2/37r - λi′ +
e2(1/r - 1/Rc)(2/ꢀs - 1/n2) (2)
where F is Faraday’s constant, e is the elementary charge, and
Eox(D) and Ered(A) are the oxidation and reduction potentials

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3617
TABLE 4: Deconvolution of Fluorescence Spectra of
D[3π3]Aa
ECT
CCT
solvent
∆f
νmax
∆ν
b
νmax
∆ν
b
cyclohexane
0.100 25.62
4.74 -0.312 19.04
4.75b -0.107 17.66
4.75b -0.092 17.53
5.56 -0.113
4.76 -0.122
4.75b -0.120
4.75b -0.119
di-n-pentyl ether 0.173 21.42
di-n-butyl ether 0.194 21.07
diisopropyl ether 0.237 19.11
diethyl ether
ethyl acetate
0.251 18.36
0.292 15.87
5.33 -0.090
5.57 -0.021
a Values for νmax and ∆ν are given in 103 cm-1
value (see text).
.
b ∆ν fixed at this
fluorescence spectroscopy (see “Time-Resolved CT Fluores-
cence” section). The fluorescence spectra of D[3π3]A in
diisopropyl ether, diethyl ether, ethyl acetate (not shown in
Figure 3A), and THF have normal band shapes and widths.
Bandwidths ∆ν1/2 range from 5.34 to 5.57 × 103 cm-1. On the
basis of the electrostatic nature of the driving force for folding,
less CCT emission is expected in these polar solvents and the
band maxima are considered to be representative for the ECT
species.
Figure 4. Plot of charge-transfer fluorescence maxima νf versus the
solvent polarity parameter ∆f: (2) D[3π3]A; (4) maxima for D[3π3]-
A obtained by deconvolution of the experimental spectrum (eq 3). Note
that these maxima have not been used in the Lippert-Mataga
analysis: (b) D[7]A; (O) maxima derived from time-resolved data for
D[7]A. The solid line is the fit for ECT emission of D[3π3]A; the
dotted line is a guide to the eye for CCT emission of D[3π3]A.
Since we were not able to distinguish between the emission
of the extended and the compact CT conformers in di-n-pentyl
ether and di-n-butyl ether, we deconvoluted43 the steady-state
spectra by skewed Gaussian functions:44
In eqs 5 and 6, which are in CGS units, c is the velocity of
light, and νf(0) the gas-phase position of νf. Furthermore, F
denotes the effective radius of the solvent cavity enclosing the
CT species.
I(ν) ) fmax exp(- ln 2[ln(1 + 2b(ν - νmax)/∆ν)/b]2) (3)
In Figure 4 the fluorescence maxima νf found for D[3π3]A
and D[7]A are plotted versus ∆f. A good fit to eq 5 is obtained
for the ECT emission maximum of D[3π3]A in cyclohexane
and the emission maxima in diisopropyl ether and more polar
solvents. The fitting parameters are given in Table 3. The value
The value for the full bandwidth at half-height ∆ν1/2 is given
by
∆ν1/2 ) (∆ν sinh b)/b
(4)
obtained for the slope -2µe /hcF3 allows µe to be determined.
2
Taking F as 0.4 times the length of the molecule,48 which is
known to be 16 Å from its X-ray structure,35 a µe value of 36
D is calculated. This corresponds to a unit charge separation
over 7.5 Å, which is in reasonable agreement with the Rc value
of 8.4 Å found in the crystal structure. It is worthwhile to
compare the slope from the Lippert-Mataga analysis with that
of D[3]A, since the ratio should reflect the ratio of Rc values.49
By fitting the values reported in ref 40 a slope of (-25.8 (
1.6) × 103 cm-1 and a νf(0) of (24.7 ( 0.3) × 103 cm-1 are
obtained for D[3]A. The ratio of the slopes thus equals 0.53
while the ratio of the Rc values equals 0.50, which is in line
with expectation. This provides strong support that in both
compounds full charge separation occurs.
The high energy maxima obtained by fitting the experimental
spectra of D[3π3]A in di-n-pentyl ether and di-n-butyl ether
(Table 4) are very close to the fitted line for the ECT species
(Figure 4), while the low energy maxima are on a straight line
with a small slope as indicated by the dotted line in Figure 4,
which is in accordance with emission from a CCT species.
If the maxima of D[7]A, except for the cyclohexane maxi-
mum, are fitted to eq 5, a slope of (-43.9 ( 5) × 103 cm-1
corresponding to a 37 D dipole moment or 7.7 Å charge
separation is obtained. These values are all in good agreement
with charge separation in an extended conformer.
The value for ∆ν was fixed at 4.75 × 103 cm-1, i.e., close to
the value obtained for the ECT and CCT bands in the
cyclohexane spectrum (Table 4 and Figure 1S). The positions
of the maxima for the ECT and CCT species obtained in this
way fit well in the solvatochromic trend obtained by means of
a Lippert-Mataga analysis (see below).45
The emission spectra of D[7]A are given in Figure 3B. It is
immediately evident that the emission spectrum of D[7]A in
cyclohexane is very broad, suggesting the presence of more than
one emitting species. It was found by Hoogesteger et al. by
means of nanosecond time-resolved emission measurements that
in cyclohexane the emission initially appears at 440 nm and
subsequently shifts to 500 nm, reflecting the conversion of an
ECT into a CCT species.18 These initial and final fluorescence
maxima are indicated in Figure 4 by open circles. It appears
that the initial maximum is red-shifted as compared to that of
the ECT species of D[3π3]A. In the more polar solvents the
emission bands have usual bandwidths [(5.1-5.7) × 103 cm-1
]
with maxima that shift with solvent polarity. It is thought that
these emission bands represent predominantly ECT emission.46
On the assumption that the excited-state dipole moment (µe)
is large as that of the ground state, the Lippert-Mataga
relationship47 (eq 5) predicts a linear relationship between the
fluorescence maximum νf of a CT state with dipole moment µe
and the solvent polarity parameter ∆f.
Transient Absorption Spectroscopy of D[3π3π3]A, D-
[7π3]A, D[3π7]A, and D[11]A. Transient absorption (TA)
spectra of D[3π3π3]A, D[7π3]A, and D[11]A were recorded
upon excitation with a nanosecond (308 nm, 7 ns fwhm) XeCl
excimer laser pulse in benzene. For D[3π7]A a 295 nm, 2.7 ns
fwhm laser pulse was used. The solubility of D[3π3π3π3]A
was too low to allow the recording of a TA spectrum.
Absorption maxima and lifetimes are presented in Table 5. In
2µe2
hcF3
νf ) νf(0) -
ꢀs - 1
∆f
(5)
(6)
n2 - 1
4n2 + 2
∆f )
-
2ꢀs + 1

3618 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
TABLE 5: Transient Absorption Maxima (nm) and
Lifetimes τ (ns) of Donor-Bridge-Acceptor Compounds in
Benzene at 20 °C
compound
λmax
D[3π3π3]A
D[7π3]A
D[3π7]A
D[11]A
340, 467
340, 480
a, 470
<5
10
10
46
340, 500
a Not measured.
all cases a broad band in the 460-500 nm range and a band at
340 nm were visible. The 460-500 nm band is attributed to
the N,N-dialkylanilino radical cation. The N,N-dimethylaniline
radical cation is known to absorb at 475 nm.50 The 340 nm
band is attributed to the dicyanovinyl radical anion.51 It is thus
concluded that for all compounds full charge separation occurs
upon photoexcitation.
Steady-State Fluorescence of D[3π3π3]A, D[7π3]A, D-
[3π7]A, D[11]A, D[3π3π3π3]A, [3π3]A, and [3π3π3]A. The
CT fluorescence maxima and quantum yields of the longer
D-B-A compounds, [3π3]A and [3π3π3]A, are given in Table
3. The fluorescence maxima and band shapes obtained for
[3π3π3]A are virtually identical to those of [3π3]A, indicating
that the positive charge in the charge-separated state [3π3π•+3]A•-
is localized on the olefinic bond adjacent to the acceptor.
Apparently, Coulombic attraction between the opposite charges
prevents the formation of the CT state [3π•+3π3]A•-.
For all D-B-A compounds the donor fluorescence is
strongly reduced in intensity. For D[3π3π3]A, D[7π3]A, D-
[3π7]A, and D[11]A in diisopropyl ether and more polar
solvents the residual percentage of local donor emission was
about 4, 6, 15, and 40%, respectively, of that of the model donor
compounds. In D[3π3π3]A, D[7π3]A, and D[3π3π3π3]A, also
the [3π•+3]A•--like emission is strongly quenched in all
solvents. Instead, for all compounds a broad structureless
fluorescence band is visible. Since this band is red-shifted as
compared to both the local donor fluorescence and the [3π•+3]-
A•--type CT fluorescence in all cases this emission stems from
the fully charge-separated state, i.e., the CT state with the
positive charge located on the anilino donor and the negative
charge on the dicyanovinyl acceptor (see TA spectroscopy
section). The quantum yield of this fluorescence is highest in
cyclohexane and is strongly reduced in the more polar solvents.
For compounds D[7π3]A and D[3π3π3π3]A in ethyl acetate
and THF, the CT fluorescence is so low in intensity that the
fluorescence spectra can be adequately described as the sum of
remaining donor and [3π•+3]A•--like fluorescence bands.
Fluorescence spectra of D[7π3]A and D[3π3π3π3]A in other
solvents could be well fitted as the sum of a donor fluorescence
spectrum (either D[3π3] or D[7]), a fluorescence spectrum of
[3π3]A or [3π3π3]A and a skewed Gaussian for the CT
emission band originating from the fully charge-separated state
(see Supporting Information for two examples). In this way it
was possible to obtain proper CT fluorescence maxima and
quantum yields.
Figure 5. Plot of charge-transfer fluorescence maxima νf versus the
solvent polarity parameter ∆f: (+) D[3π3π3]A, (0) D[7π3]A, (×) D-
[3π7]A, (O) D[11]A. Top line: fit for D[7π3]A. Bottom line: fit for
D[3π3π3]A.
fluorescence from a folded (CCT) species is observed, in which
the donor and acceptor are in close contact (see also TRMC
section).
Due to its low solubility, fluorescence spectra of D[3π3π3π3]-
A could only be obtained in three solvents. In addition to the
[3π•+3]A•--like emission at 415 nm, in benzene a CT emission
at 577 nm belonging to a fully charge-separated state could be
identified. The fact that its band maximum is almost identical
to that of D[3π3π3]A leaves no doubt about the folded nature
of the emitting species. It is conspicuous that the fluorescence
maxima of D[3π3π3]A, D[7π3]A, and the single maximum
obtained for D[3π3π3π3]A are systematically red-shifted in
comparison to the maxima of D[3π7]A and D[11]A (for the
latter with the exception of cyclohexane) by an average of (7.0
( 0.8) × 102 cm-1 (0.09 ( 0.01 eV, Figure 5). In other words,
the transition energy is smaller when a double bond is attached
to the ring bearing the acceptor. According to eq 2 there are
two possible explanations for this difference. The first is a
difference in reduction potential of the acceptor. However, cyclic
voltammetry revealed that the onsets of the irreversible reduction
waves of [3π3]A and [7]A do not differ by more than 0.02 V,
which means that the reduction potentials are equal within
experimental error. The second possibility is a difference in
reorganization energy λi′ upon charge recombination in the
folded state. Apparently, the nature of the bond connecting the
ring bearing the acceptor to the other part of the molecule is of
interest for the reorganization energy of the CCT species,
whereas the bond type near the donor site does not have an
effect (see below).
Time-Resolved CT Fluorescence. Charge-transfer fluores-
cence decay times τf for all D-B-A compounds in a few
solvents are given in Table 6. Lifetimes were measured with a
streak camera system upon 295 nm excitation (fwhm 2.7 ns).
To obtain a better time-resolution, in benzene and cyclohexane
additional SPC measurements were performed (excitation 312-
315 nm, fwhm instrument response time ∼17 ps).
When the CT fluorescence maxima are plotted against ∆f, a
linear relationship is found for D[3π3π3]A, D[7π3]A, D[3π7]-
A, and D[11]A (Figure 5).52 The slopes obtained for these four
compounds are virtually identical and are smaller than those of
D[3π3]A, D[7]A, and even D[3]A. When the distance between
donor and acceptor Rc is estimated on the basis that the ratio of
the Lippert-Mataga slopes is essentially equal to the ratio of
the Rc distances (using a slope of 25.8 × 103 cm-1, with Rc )
4.16 Å for D[3]A), extremely small values of 2.0-2.2 Å are
obtained.53 This leads to the conclusion that in all cases
In steady-state measurements on D[3π3]A, two fluorescence
bands were observed in cyclohexane and benzene (see “Solvent
Dependence of the Steady-State Fluorescence of D[3π3]A and
D[7]A” section) of which one was assigned to an ECT species
and the other to a CCT species. It was noted previously18 that
the CCT fluorescence of D[3π3]A in cyclohexane seems to
appear instantaneously upon excitation. This was tentatively
attributed to the occurrence of charge separation in a partly
folded conformation or to very rapid folding of a stretched

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3619
TABLE 6: Charge-Transfer Fluorescence Rise Times τi and Decay Times τf (in ns) of Donor-Bridge-Acceptor Compounds in
Various Solvents at 20 °C. Data Determined near the Maximum of the Fluorescence Band (See Table 3)
D[3π3]A
D[7]A
D[3π3π3]A
D[7π3]A
D[3π7]A
D[11]A
τi τf
solvent
τi
τf
τi
τf
τi
τf
τi
τf
τi
τf
cyclohexane
di-n-pentyl ether
benzene
di-n-butyl ether
diisopropyl ether
a
0.24 (ECT); 12.6 (20b)
1.3
13 (24b)
a
20.5
1.4
13.6
0.9
22.9
c
21
20
31
19
19
a
0.36 (ECT); 1.6 (2b)
a
19 (36b)
29
14
0.21
2.7
2.7
0.8
4.3
4.0
0.7
10.9
6.4
10.5
a
∼1.6
a Rise of the fluorescence signal not observed. b Values from ref 18. c Not observed as a consequence of too low intensity.
TABLE 7: Fluorescence Lifetimes (in ns) [Quantum
TABLE 8: TRMC Data for Donor-Bridge-Acceptor
Compounds
Yields]18 of Donor and Acceptor Chromophores at 20 °C
solvent
D[3π3]
D[7]
[3π3]A
[3π3π3]A
compound
τ (ns)
µ2/θa (D2 ps-1
)
µ
cyl (D)
µ
sph (D)
cyclohexane
benzene
THF
1.87 [0.25]
2.66 [0.34]
2.50
1.82 [0.25]
2.56 [0.35]
2.43
D[3π3]A18
2.0
36.0
2.2
5.0
6.5
25
3.43
3.74
2.63
2.50
4.73
5.17
2.63
27.9
29.4
32.3
31.6
43.6
45.5
40.7
16.8
17.6
16.4
15.7
21.8
22.8
18.2
1.32
2.07
1.50
2.31
D[7]A18
D[3π3π3]A
D[7π3]A
D[3π7]A
D[11]A
conformation. We have now determined the lifetime of the ECT
state to be 0.24 ns in cyclohexane and 0.36 ns in benzene and
confirmed the instantaneous appearance of CCT fluorescence.
This means that the decay of the ECT fluorescence is not
coupled to the appearance of the CCT fluorescence. This
observation shows that different conformers of the fully charge-
separated state are present directly after photoexcitation followed
by charge separation. The ECT species, for instance, could be
the anti conformer while the already partly folded syn conformer
could be held responsible for the apparent instant appearance
of CCT fluorescence.
D[3π3π3π3]A
2.2
a Values are corrected for the quantum yield for charge separation
as obtained from time-resolved fluorescence measurements (Table
9).
for a trichromophoric compound consisting of a 1,3-diphenyl-
propanedioate boron oxalate acceptor, a 4,4′-biphenylene D1
donor and a methoxynaphthalene D2 donor. A 3- to 4-fold
increase in the charge-recombination rate upon addition of two
methoxy groups to a benzene spacer in a trichromophoric
compound with a zinc porphyrin donor and a 1,4-naphtho-
quinone acceptor has been observed by Wasielewski et al.57 In
this case a virtual D2-D1•+-A•- state was held responsible for
the increased charge-recombination rate.
Time-Resolved Microwave Conductivity (TRMC). Using
the TRMC technique, the change in microwave conductivity
∆σ occurring on flash-photolysis at 308 nm of benzene solutions
of the present compounds was measured. To a good approxima-
tion ∆σ is related to the concentration of excited states formed
Ne, the excited-state dipole moment µe, and the dipole relaxation
time θ via58,59
In cyclohexane the lifetimes of the CCT species of the
D-B-A compounds are in the range 13-23 ns. Upon increas-
ing the solvent polarity, the CCT lifetimes of D[3π3]A,
D[3π3π3]A, and D[7π3]A decrease dramatically, while for D-
[7]A and D[11]A, containing fully saturated bridges, this is not
the case. For example, in benzene, which is considered to be a
quasi polar solvent (as invoked from CT emission maxima),
the lifetime of the CCT state of D[3π3π3]A is approximately
10 times shorter than that of D[11]A. For D[3π7]A intermediate
behavior is observed. By comparing the CCT lifetimes of D-
[3π3π3]A and D[7π3]A it appears that the extra olefinic bond
next to the donor in D[3π3π3]A has only a very small
accelerating effect on the recombination rate (factor of about
1.5). These observations indicate that the olefinic bond provides
an additional decay channel and that this channel becomes less
important as the distance between the olefinic bond and the
acceptor is enlarged. This is quite remarkable since the
compounds are in a folded conformation. The most probable
decay channel is governed by an increased coupling of the CCT
state with the [3π•+3]A•--like “local” CT state. This state has
a short lifetime (see Table 7) in all solvents employed and its
energy drops faster with increasing solvent polarity than that
of the CCT states (see the slopes of Lippert-Mataga plots in
Table 3).54 Thus, the energetic distance between the [3π•+3]A•-
state and the CCT state decreases upon increasing solvent
polarity by which the coupling between these states is enlarged.
A small degree of admixture of a short-lived state can strongly
reduce the lifetime of a CT state.55 If the degree of mixing is
small, the magnitude of the CT state’s dipole moment is not
substantially affected. For D[3π7]A it is anticipated that the
D[3π•+7]A•- state lies higher in energy than the D[3π3π•+3]A•-
state of D[3π3π3]A due to the larger distance between the
charges. This state could provide the additional decay channel
in D[3π7]A.
(n2 + 2)2Ne µe2
∆σ )
·
(7)
27kBT
θ
From convolution fits to the TRMC transients, the excited-
state lifetime τe and the parameter µe /θ can be derived, and
these are given in Table 8. In deriving µe /θ, the fact that the
2
2
quantum yields of formation of the dipolar excited states of the
longer-bridge compounds are somewhat smaller than unity as
derived from the optical data, was taken into account (Table
9).
The TRMC results confirm the highly dipolar, charge-transfer
nature of the relaxed S1 states of all of the D-B-A compounds
investigated. The decay times of the TRMC transients are in
reasonable agreement with the decay times found for the
transient absorption and the fluorescence of the same compounds
in benzene (Tables 5 and 6). Of particular interest is the
influence of the bridging unit on the decay time since this is
governed by charge recombination involving electron transfer
between the donor and acceptor moieties.
For D[7]A and D[3π3]A, both of which have seven carbon-
carbon bonds between donor and acceptor and close to equal
donor-acceptor distances in the ground state, the recombination
time is more than an order of magnitude shorter for the latter
Unexpectedly fast charge recombination possibly due to an
intermediate D2-D1•+-A•- state has been observed before56

3620 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
TABLE 9: Donor Fluorescence Lifetimes (τ in ns) and Rates of Photoinduced Intramolecular ET (ket in s-1 [Calculated
Quantum Yield of Charge Separation Φet]) of Donor-Bridge-Acceptor Compounds at 20 °C
D[3π3π3]A
D[7π3]A
ket/109
D[3π7]A
ket/109
D[11]A
ket/109
D[3π3π3π3]A
τ
ket/109
solvent
τ
ket/109
τ
τ
τ
cyclohexane
benzene
THF
0.17
0.11
0.19
5.3 [0.91]
8.7 [0.96]
4.9 [0.92]
0.30
0.21
0.36
2.8 [0.84]
4.4 [0.92]
2.4 [0.85]
0.20
0.18
0.38
4.5 [0.89]
5.2 [0.93]
2.2 [0.85]
0.46
0.59
0.92
1.6 [0.75]
1.3 [0.77]
0.68 [0.62]
a
a
0.93, 1.5
1.06, 1.92
0.70 [0.65]
0.54 [0.58]
a Solubility too low.
compound. The presence of a central π-bond in the bridge would
therefore appear to have a marked accelerating effect on electron
transfer.
maxima of D[3π3π3]A and D[7π3]A as compared to those of
D[3π7]A and D[11]A which indicate that their CCT state is
stabilized.
Energetics of Charge Separation. An important quantity
in the analysis of the charge-separation process is the Gibbs
energy difference ∆G between the initial and final states of the
charge-separation step D•+-B-A•- r D*-B-A. A well-
known way to estimate ∆G is by means of the Weller
equation.60,61 The usefulness of the Weller equation critically
depends on the choice of the mean ion radius r and the donor
acceptor distance Rc. A more direct way to obtain an estimate
for ∆G is possible when both charge-transfer absorption and
emission bands can be measured. The energy of a CT state can
then be estimated by the mean value of the absorption and
emission maxima.42 For most of the D-B-A compounds in
the present study these values cannot be obtained, either because
CT absorptions involving full charge separation are not discern-
ible in the spectra or because of overlap with other bands. For
D[3π3]A in cyclohexane, however, the onset of a CT absorption
band and the emission from an ECT species in cyclohexane
could be determined. Although a maximum of the CT absorption
band is not present, the intersection of the onsets of both bands
also yields E00. Thus, using the spectra represented in a reduced
form62 and scaled in intensity in such a way that the mirror
image relationship holds for the onsets of both bands, the E00
value for the CT transition in cyclohexane is estimated as 3.52
eV. For D[7] the E00 value obtained in this way in cyclohexane
is 3.92 eV, so that ∆G ∼ -0.40 eV for charge separation in
D[3π3]A in cyclohexane. This value is in reasonable agreement
with our previous18 estimate of -0.52 eV which was obtained
by using the Weller equation. For the shorter homologue D[3]A
using the CT absorption and fluorescence maxima hνa ) 3.63
eV (342 nm) and hνf ) 2.77 eV (22.3 × 103 cm-1) in n-hexane40
a value for ∆G of ∼ -0.72 eV is calculated. When obtained in
a nonpolar solvent for which the outer or solvent reorganization
energy can be neglected,63 CT absorption and fluorescence
maxima also give access to the inner reorganization term λi via2
For the four compounds with a total of 11 carbon-carbon
bonds in the bridge, D[11]A, D[3π7]A, D[7π3]A, and D[3π3π3]-
A, the lifetime toward recombination is found to decrease by a
factor of approximately 5 in going from the completely saturated
bridge to bridges containing one π-bond. A further decrease in
lifetime by a factor of 2-3 is found on introducing a second
π-bond in the bridge. The present results on both the seven-
and 11-bond D-B-A compounds therefore clearly demonstrate
the pronounced positive influence on charge recombination of
the substitution of a π- for a σ-bond in the bridging unit. While
we cannot exclude the possibility that the effects observed arise
to a certain extent from the greater conformational flexibility
of the bridges containing double bonds, we consider that this is
insufficient to explain the results completely. We conclude that
the intervening double bonds have a super-exchange function
and facilitate charge recombination via electron transfer through
the bridge.
As discussed in previous sections, there is evidence that the
geometrical conformation of the molecules in the excited state
may differ considerably from that in the ground state. Because
of this it is not possible to make a reasonable estimate of the
rotational relaxation time θ, which is required to derive an
estimate of the dipole moment of the relaxed S1 state from the
values of µ2/θ determined experimentally. In Table 8 we
therefore give two estimates for µe, one of which, µcyl, is based
on a value of θ estimated for an outstretched, cylindrical
geometry, and the other, µsph, is based on a compact spherical
geometry.59
The value of µsph corresponds to the smallest dipole moment
possible for S1. For all D-B-A compounds this is significantly
larger than the value of 10 to 12 D associated with close-contact,
donor-acceptor exciplexes. On the other hand the values of
µ
cyl are in general smaller than would be expected for complete
charge separation between donor and acceptor moieties in the
ground-state geometry, i.e., ca. 40 and 60 D for the seven- and
11-bond bridged compounds, respectively. The TRMC results
are therefore in general agreement with the previous conclusions
that a substantial change in the molecular geometry occurs after
charge separation which results in a decrease in the distance
between the donor and acceptor moieties. Complete folding of
the bridge, resulting in the formation of a contact donor-
acceptor pair does not however occur for any of the compounds
studied.
hνa - hνf ) 2λi
(8)
Upon use of the same data, for D[3]A a value for λi of
approximately 0.4 eV is obtained. This value is expected not
to depend much on the donor-acceptor separation distance Rc,
since the main structural reorganization upon charge transfer
takes place on the donor and acceptor sites. Thus, assuming
that λi = 0.4 is also valid for D[3π3]A and D[7]A, it can be
concluded that for these compounds (in cyclohexane) charge
transfer takes place near the optimal region where -∆G ) λ.
For the longer homologues ET in cyclohexane will almost
certainly take place in the normal region for which -∆G < λ
since the longer donor-acceptor distance Rc for these com-
pounds will decrease the driving force for charge separation. It
is known that with increasing medium polarity, both the driving
force and the total reorganization energy λ increase, but since
these contributions tend to cancel each other2 it is expected that
It is conspicuous that the µe2/θ data divide the four 11-bonds-
bridged compounds into two pairs. The values for the com-
pounds with an olefinic bond next to the acceptor are much
smaller than those for the compounds with a single bond at the
corresponding position. This could well be related to the
occurrence of different structures of the two sets of compounds
(which affects θ rather than µe, since the Lippert-Mataga slopes
are similar). This is supported by the red-shifted fluorescence

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3621
the above conclusions will hold throughout the applied solvent
polarity range.
In the following analysis a survey of the influence of bridge
topology on the rate of ET ket is made. Although formally
incorrect, it will be assumed that the ET rate processes of the
different conformers can be represented by a single rate constant.
For D[3π3π3]A, D[3π7]A, D[7π3]A, and D[11]A the electron
transfer was found to occur in an extended conformation prior
to folding. In that case the electronic coupling energy between
reactant and product states Hrp is expected to be small so that
the ET process is nonadiabatic and can be described by
semiclassical electron transfer theory.64 The rate constant ket is
then given by63
Bridge Dependence of the Charge-Separation Process. To
compare the charge-separation dynamics as a function of bridge
topology, we determined the fluorescence lifetimes of the model
donors and acceptors (Table 7). The emission could in all cases
be fitted well with a single exponential. In addition, the decay
of the residual local donor emission of the D-B-A compounds
was measured (Table 9). For D[3π3]A and D[7]A no reliable
decay data could be obtained. In the case of compounds D-
[3π3π3]A, D[7π3]A, D[3π7]A, and D[11]A, the residual local
donor fluorescence could be satisfactorily fitted with two
lifetimes: a very short component and a component with a
lifetime in the order of those of the model donor compounds
D[3π3] and D[7]. For D[3π3π3π3]A more exponentials were
necessary. The component comparable to the lifetime of the
model donor compounds probably stems from minute amounts
of molecules which lack the acceptor. When it is assumed that
the short component of the donor fluorescence decay is
shortened only by the additional pathway of ET and that the
latter process is irreversible, rates of intramolecular electron
transfer ket can be calculated with
4π2
h
ket )
Hrp2FCWD
(11)
where FCWD is the Franck-Condon weighted density of states.
If it is also assumed that compounds of comparable length and
identical donor-acceptor pair such as used here possess a similar
FCWD value, differences in ket will directly reflect differences
in Hrp2. Since Hrp is predominantly determined by the nature of
the bridge, ket also directly gives the effect of incorporation of
an olefinic bond in the bridge.
The following trend is observed: the rates for D[3π3π3]A
are roughly 2 times faster than those for D[7π3]A and D[3π7]-
A, which are of similar magnitude. The rates for D[7π3]A and
D[3π7]A in turn are about 3 times faster than those for D[11]-
A. For the quantum yield of charge separation also the trend
D[3π3π3]A > D[7π3]A ∼ D[3π7]A > D[11]A is found.
Although this suggests that replacement of the first single bond
by a double bond has a larger effect than replacement of the
second, on average an increase in the rate of charge separation
of 3.0 ( 0.8 per olefinic bond is observed. This corresponds to
an increase of the electronic coupling of 1.7 ( 0.2 per bond.
At this point it is of interest to note that the introduction of a
double bond increases the electronic coupling in two hardly
separable ways. It increases the donor-acceptor interaction since
the bridge facilitates transport of electronic effects. Concurrently,
the donor-acceptor distance is reduced, as a CdC double bond
is shorter than a C-C single bond. The shorter distance is
expected to affect ∆G, λ and FCWD, which is a function of
both ∆G and λ. As it is likely that ket follows the usual distance
dependence ket exp(-âRc), the net effect will be that upon
going from D[11]A to D[3π3π3]A ket also increases as a
consequence of the shorter donor-acceptor distance. This makes
the factor of 1.7 ( 0.2 an upper limit to the purely electronic
coupling. When viewed in terms of the number of bonds, one
may however also state that the effect on the electronic coupling
is a factor of 1.7 ( 0.2 per double bond.
An effect of the incorporation of two olefinic bonds on
the rate of charge separation has been observed by de Rege
et al.65 as well. They found that incorporation of two double
bonds in a saturated bicyclic hydrocarbon bridge enhances the
rate of charge separation by a factor of 2. Introduction of
exocyclic double bonds in the oligo(cyclohexyl) bridge thus has
a larger effect. This larger effect might find its origin in the
favorable interaction between the double bond and Hax-C-
C-Hax units.12,13
Solvent Dependence of the Rate of Charge Transfer in
D[11]A. To obtain an indication of the magnitude of the barrier
for ET in D[11]A (for which the donor-acceptor separation
distance is considered to be well-defined) the donor fluorescence
lifetimes of this compound and those of D[7] were measured
in six solvents (Table 10). ET rates obtained from these data
via eq 9 range from 0.68 × 109 s-1 in THF to 1.6 × 109 s-1 in
cyclohexane. Hence, the effect of the medium polarity on ket is
1
1
ket )
-
(9)
τf(DBA) τf(D)
where τf(DBA) is the donor fluorescence lifetime in the
D-B-A compound and τf(D) the fluorescence lifetime of the
appropriate model donor compound (D[3π3] or D[7]). Within
the same approximation the quantum yield of ET Φet can be
calculated with
ket
ket
Φet )
)
)
kf(D) + knr(D) + ket 1/τf(D) + ket
τf(D) - τf(DBA)
τf(D)
(10)
where kf(D) and knr(D) are the fluorescent and nonradiative
decay rates of the model donor compound, respectively.
Obtained values of ket and Φet are also collected in Table 9.
With the SPC setup the rise of the CT fluorescence of D-
[3π3π3]A, D[7π3]A, and D[3π7]A in cyclohexane and benzene
was observed, with time constants in the range 0.2-1.4 ns.
Unfortunately, the CT fluorescence band of D[11]A was very
weak which, in combination with its low solubility, hampered
a reliable study of the possible rise of CT fluorescence. When
the lifetimes of donor fluorescence of D[3π3π3]A, D[3π7]A,
D[7π3]A, and D[11]A are compared to the time constants for
the ingrowth of the CCT fluorescence (Table 6) it can be seen
that in all cases where both values are available the former are
substantially shorter. In other words, the decay of the local donor
fluorescence does not reflect the formation of the CCT state.
An intermediate state has to be present and this is supposed to
be an ECT state. Hence, ET in an extended conformer occurs
which is followed by folding to a compact or CCT state. This
sequence is known as “harpooning”.20-23 For D[7]A a harpoon-
ing mechanism is operative as well, but in the case of D[3π3]-
A ingrowth of the CCT emission could not be detected whereas
decay of the ECT emission was observed (see above). This can
be explained by assuming different behavior for the syn and
anti conformers of D[3π3]A. ECT emission may only be
produced by the anti conformer whereas the rapid appearance
of CCT emission may originate from contraction of syn
conformers. The latter are already partly folded in the ground
state.

3622 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
TABLE 10: Donor Fluorescence Lifetimes (τ in ns) of D[7]
and D[11]A and Rates of Photoinduced Intramolecular ET
[ket in s-1 [Calculated Quantum Yield of Charge Separation
Φet]] for D[11]A in Various Solvents at 20 °C
Experimental Section
General. All reactions and distillations were carried out under
an atmosphere of dry N2 unless stated otherwise. Commercially
available reagents were used without further purification. THF
and diethyl ether were distilled from Na/benzophenone prior
to use. MeCN was distilled from CaH2. EtOH was stored on
molecular sieves (3 Å). N,N,N′,N′-Tetramethylethylenediamine
(TMEDA) was dried by boiling with LiAlH4 under reflux for 3
h, followed by distillation under reduced pressure (0.01 Torr).67
Column chromatography was performed using ACROS silicagel
0.035-0.070 mm, pore diameter ca. 6 nm. Thin-layer chroma-
tography was performed on Merck silicagel 60 F254. Spots were
detected by the use of iodine vapor and/or UV light. Melting
points were determined on a homemade melting point apparatus
or on a Mettler FP5/FP51 photoelectric melting point apparatus
and are uncorrected. Gas chromatography was performed on a
Varian 3350 instrument equipped with a DB-5 (i.d. 0.309 mm,
25 m) capillary liquid-phase column and an FID using N2 as
carrier gas. NMR spectra were recorded on a Bruker AC 300
spectrometer operating at 300.13 MHz for 1H NMR and at 75.47
MHz for 13C NMR. Samples were dissolved in deuterated
chloroform unless stated otherwise. Chemical shifts (in ppm)
D[11]A
D[7]
solvent
τ
τ
ket/109
cyclohexane
di-n-pentyl ether
di-n-butyl ether
diisopropyl ether
THF
1.82
2.1
2.10
1.99
2.43
2.56
0.46, 1.6
0.73, 2.2
0.84, 2.0
0.75, 1.8
0.92, 1.9
0.59, 2.7
1.6 [0.75]
0.89 [0.65]
0.71 [0.60]
0.83 [0.62]
0.68 [0.62]
1.3 [0.77]
benzene
very moderate. This indicates that in D[11]A the donor-
acceptor distance is such that the barrier for ET is of similar
magnitude in all solvents. According to design criteria for ET
systems,66 this near solvent independence suggests that the
donor-acceptor combination, along with the bridge, provides
an efficient ET system. This boils down to the fact that the
quantity Eox(D) - Ered(A) - E00 will not deviate more from
the optimal value than ca. 0.5 eV. A glance at Table 9 indicates
that this is valid for the whole series of compounds D[3π3π3]-
A, D[7π3]A, D[3π7]A, and D[11]A. Note that for D[3π3]A
and D[7]A it was inferred that ET will be close to barrierless,
since it was found that -∆G ∼ λ.
1
are given relative to internal TMS (0 ppm) in the case of H
NMR and relative to CDCl3 (77.00 ppm) in the case of 13C
NMR. Infrared spectra were recorded on a Mattson Galaxy
Series FTIR 5000 operating with 2 cm-1 resolution. Solids were
measured in KBr pellets, while liquid materials were measured
as a thin film between NaCl plates. Peak maxima are given in
cm-1, while intensities are designated as s (strong), m (medium),
or w (weak). Molecular mechanics (MM2) calculations were
performed using the MM2 force field as implemented in
CambridgeSoft Chem3D Pro, version 4.0 (1997). Cyclic vol-
tammetry was performed with a EG&G Princeton Applied
Research model 263A potentiostat/galvanostat. The scan rate
employed was 50 mV s-1. Measurements were conducted in
MeCN (freshly distilled from CaH2) using anhydrous tetrabu-
tylammonium hexafluorophosphate in 0.1 M concentration as
the electrolyte. The solution was purged with N2 prior to the
measurement. A three-electrode setup was used, with a Pt disk
electrode (area 0.78 mm2), a Pt counter electrode and a Ag/
AgNO3 (0.1 M in MeCN) reference electrode. Solute concentra-
tions were 2 mM. MALDI TOF-MS (positive-ion mode) was
performed on a Perkin-Elmer/PerSeptive Biosystems Voyager-
DE-RP MALDI TOFF-MS [N2 laser; λexc ) 337 nm (3 ns
pulses)]. TOF-MS spectra were recorded in the reflectron mode.
The samples D[3π3π3π3]A and D[3π7]A were mixed with
dihydroxybenzoic acid solution (3 mg mL-1); 1 µL of the
suspension was loaded on the Au-sample plate.
Conclusions
Photoinduced ET takes place in extended conformers of D-
[7]A, D[3π3π3]A, D[7π3]A, D[3π7]A, D[11]A, and D[3π3π3π3]-
A producing a fully charge-separated state. The ET step is
followed by folding, even in solvents as polar as diethyl ether.
For D[3π3]A both the extended and folded fully charge-
separated CT conformers were detected, but the decay of the
ECT fluorescence is not coupled to the ingrowth of the CCT
fluorescence. This is explained by the presence of both syn and
anti conformers in solution, which exhibit markedly different
excited-state behavior.
The presence of an olefinic bond in the bridge influences
the charge-separation and -recombination rates, as well as the
excited-state properties of the studied D-B-A compounds
considerably. The fluorescence maxima of CCT states with a
[3π3]A-type acceptor are consistently 0.09 eV red-shifted as
compared to those of compounds with a [7]A-type acceptor.
This might be explained by a structural difference in the two
sets of CCT conformers. TRMC measurements also point to a
structural difference in the CCT species.
In the series D[3π3π3]A, D[3π7]A, D[7π3]A, and D[11]A,
the rate of charge separation D•+-B-A•- r D*-B-A
increases with a factor of 3.0 ( 0.8 per olefinic bond and is
not strongly affected by the position of the olefinic bond in the
bridge. Since the incorporation of an olefinic bond also slightly
reduces the donor-acceptor distance, this corresponds to a
maximum factor of 1.7 ( 0.2 for the increase of the electronic
coupling with the introduction of an olefinic bond. The charge-
recombination rate of the CCT state increases with increasing
solvent polarity for compounds whose bridge contains an
olefinic bond. This behavior is explained by a recombination
process that involves D-B•+-A•- states. The effect is most
pronounced in case the olefinic bond is located next to the
acceptor, i.e., for compounds with a [3π3]A-like acceptor. Thus,
in benzene the recombination rate for D[3π3π3]A is 11 times
that found for D[11]A. Hence, although incorporation of a
double bond accelerates the charge-separation process, the effect
on the charge-recombination rate is even more pronounced.
Steady-State Absorption and Fluorescence Spectroscopy.
Details of the instrumentation used to record absorption and
fluorescence spectra are given in ref 34. Fluorescence quantum
yields were determined relative to quinine sulfate in 0.5 M H2-
SO4 (φ ) 0.546)68 at 20 °C (excitation at 300 nm). To accurately
measure the low (φ < 0.05) quantum yields of CT emission a
solution of quinine sulfate with known absorbance at the
excitation wavelength was diluted by a known factor in such a
way that the maximum emission intensity of the quinine sulfate
solution was comparable to that of the sample. To correct for
the wavelength dependence of the fluorescence intensity due
to the use of monochromators, the fluorescence intensity I(λ)
was multiplied by λ2 when converting the spectra to an energy
scale. Solutions were purged with Ar for at least 10 min prior
to measurement. Corrections for solvent refractive indices were
made according to Eaton68 when needed.

Photoinduced Intramolecular Charge Separation
J. Phys. Chem. A, Vol. 107, No. 19, 2003 3623
Spectroscopic grade solvents were used throughout with the
following exceptions; diisopropyl ether (ACROS 99+ %) and
di-n-butyl ether (Fluka, p.a.; low in aromatic compounds;
g99.5%) were stirred with LiAlH4 for at least 1 day and then
distilled. Di-n-pentyl ether (Fluka, 99%) was purified as follows.
Five parts of the ether were stirred with one part of sulfuric
acid (Merck, p.a. 95-97%) for 1 day. The brown mixture was
poured into 2.5 parts of water and the layers were separated.
The organic layer was washed with water, saturated NaHCO3
solution and water (similar volumes), dried on MgSO4, and
filtered. This procedure was repeated until a brown color did
not develop anymore upon stirring with sulfuric acid. Subse-
quently, the ether was treated as described for diisopropyl ether.
Ethyl acetate (ACROS, HPLC grade, g99.5%) was stirred with
CaH2 for 3 days and subsequently distilled. THF and diethyl
ether were distilled from Na/benzophenone. All purified solvents
were checked for spurious emission.
4-[4-(3,3-Dimethyl-1,5-dioxaspiro[5.5]undecan-9-ylidene)-
cyclohexylidene]-1-phenylpiperidine (4). To a suspension of
â-hydroxy acid 3 (2.35 g, 4.86 mmol) in MeCN (50 mL) was
added N,N-dimethylformamide dineopentyl acetal (2.35 g, 10.2
mmol) and after stirring for 1 h at room temperature the reaction
mixture was heated to reflux overnight. The resulting yellow
solution was cooled to -20 °C and the resulting precipitate was
filtered off, washed with cold MeCN, and dried to yield an off-
white solid (1.73 g, 85%): mp 222 °C (dec). 1H NMR δ: 0.98
(s, 6H), 1.79-1.83 (m, 4H), 2.23-2.30 (m, 12H), 2.44-2.48
(m, 4H), 3.20-3.24 (m, 4H), 3.53 (s, 4H), 6.78-6.82 (m, 1H),
6.91-6.93 (m, 2H), 7.22-7.25 (m, 2H). 13C NMR δ: 22.8,
25.1, 29.0, 29.4 (2 ×), 30.2, 33.4, 50.4, 70.1, 97.8, 115.9, 118.8,
125.6, 128.2, 129.0 (2 ×), 130.1, 151.4. FT-IR ν˜max: 3096, 3067,
3040, 3023, 2976, 2953, 2928, 2890, 2866, 2839, 2830, 1599,
1462, 1445, 1429, 1113, 1099, 748, 685.
4′-(1-Phenylpiperidin-4-ylidene)-1,1′-bi(cyclohexyliden)-4-
one (5). Acetal 4 (1.72 g, 4.09 mmol) was dissolved in THF
(250 mL), and 5% HCl solution (25 mL) was added. The
mixture was heated to reflux for 4 h. After cooling to room-
temperature THF was removed under reduced pressure and the
resulting suspension was extracted with CHCl3 (2 × 150 mL).
The combined organic extracts were washed with a saturated
NaHCO3 solution and water. Drying over MgSO4, followed by
evaporation of the solvent under reduced pressure, gave an off-
white solid (1.25 g, 91%): mp 155 °C (dec). 1H NMR δ: 2.32
(s, 8H), 2.39-2.44 (m, 4H), 2.45-2.49 (m, 4H), 2.55-2.59
(m, 4H), 3.21-3.25 (m, 4H), 6.79-6.84 (m, 1H), 6.91-6.94
(m, 2H), 7.23-7.28 (m, 2H). 13C NMR δ: 26.5, 28.8, 28.9,
29.3, 40.5, 50.4, 115.9, 118.9, 124.3, 126.1, 129.0, 129.4, 131.9,
151.3, 212.9. FT-IR ν˜max: 3092, 3063, 3040, 3017, 2971, 2959,
2890, 2841, 2828, 2810, 1723, 1599, 1460, 1438, 1423, 758,
696.
Time-Resolved Measurements. Picosecond single-photon-
counting (SPC) fluorescence decay traces were measured and
analyzed using the setup and computer programs described in
ref 69. Samples were degassed by purging with Ar for 10 to 15
min and were excited at 310-320 nm. Time-resolved micro-
wave conductivity was performed as described in ref 18.
Transient absorption spectroscopy of D[3π3π3]A, D[7π3]A, and
D[11]A was performed using the equipment and procedures
described in ref 18. In the case of D[3π7]A the equipment and
procedures described in ref 34 were used.
Synthesis. The syntheses of D[3π3]A, D[7]A, D[3π3], D-
[7], and [3π3]A have been described in ref 18. The syntheses
of [3π3]A and [3π7]A have been described in ref 70. D[3π3]A
was exhaustively purified by column chromatography with CH2-
Cl2 as eluent followed by two crystallizations from ethyl acetate
(6.5 mg mL-1, reflux until reaching room temperature) under a
N2 atmosphere. The purity was monitored by recording the
emission spectrum in THF, the relative intensity of “local”
emission functioning as an indicator for the amount of impurity.
D[7]A was purified by column chromatography with CH2Cl2
as eluent. The synthetic procedures for the preparation of 2e
and D-B-A compounds D[3π3π3π3]A, D[7π3]A, D[11]A,
and D[3π7]A are available as Supporting Information.
Synthesis of D-B-A Compounds. 9-[1-Hydroxy-4-(1-
phenylpiperidin-4-ylidene)cyclohexyl]-3,3-dimethyl-1,5-
dioxaspiro[5.5]undecane-9-carboxylic acid (3). A two-necked
flask was filled with THF (50 mL) and diisopropylamine (2.17
g, 21.7 mmol). To this solution n-butyllithium in n-hexane (14.5
mL of a 1.46 M solution, 21.2 mmol) was added at -40 °C.
After the mixture was stirred for 30 min, 3,3-dimethyl-1,5-
dioxaspiro[5.5]undecane-9-carboxylic acid (2e)11 (2.37 g, 10.4
mmol) in THF (50 mL) was added at -40 °C. The reaction
mixture was stirred at 50 °C for two h and recooled to -40 °C.
The compound 1-phenylpiperidin-4-one (19)40 (2.63 g, 10.3
mmol) in THF (50 mL) was added, and the reaction mixture
was stirred at 50 °C for another 2 h. After the mixture was
cooled to room temperature, the resulting suspension was poured
on ice (150 g) and diluted with diethyl ether (200 mL). The
layers were separated, and the water layer was washed with
diethyl ether (2 × 100 mL). A white precipitate formed in the
water layer which was filtered off and dried in vacuo over KOH.
The white solid (lithium salt of 3) was then suspended in water
(50 mL) and acidified to pH 1 using 3 M HCl solution. The
suspension was stirred for 90 min, and the white solid was
filtered off and dried in vacuo over KOH, yielding 3 (2.35 g,
47%): mp 210 °C (dec). FT-IR ν˜max: 3434, 2971, 2955, 2911,
2843, 2900-2300, 1672, 1493, 1462, 1438, 1119, 770, 700.
[4′-(1-Phenylpiperidin-4-ylidene)-1,1′-bi(cyclohexyliden)-
4-ylidene]malononitrile (D[3π3π3]A). A mixture of ketone 5
(0.63 g, 1.88 mmol), malononitrile (0.24 g, 3.64 mmol),
ammonium acetate (0.19 g, 2.5 mmol) and acetic acid (0.45
mL, 7.9 mmol) in benzene (130 mL) was refluxed for 2 h in a
Dean-Stark apparatus. After ca. 100 mL of benzene-water
mixture was distilled off, the remaining suspension was cooled
to room temperature. The greenish precipitate formed was
filtered off, washed with benzene and purified by column
chromatography (silica; eluent CHCl3). This gave D[3π3π3]A
1
(0.41 g, 57%) as a greenish solid: mp 214 °C. H NMR δ:
2.30 (s, 8H), 2.46-2.50 (m, 8H), 2.73-2.77 (m, 4H), 3.21-
3.24 (m, 4H), 6.79-6.84 (m, 1H), 6.91-6.94 (m, 2H), 7.23-
7.26 (m, 2H). 13C NMR δ: 28.2, 28.8, 29.0, 29.4, 34.5, 50.4,
83.0, 111.6, 115.9, 118.9, 123.4, 126.5, 129.0, 129.1, 133.2,
151.3, 184.9. FT-IR ν˜max: 3090, 3077, 2982, 2969, 2903, 2839,
2230, 1599, 1504, 1465, 1439, 1425, 750, 688. Anal. Calcd for
C26H29N3: C, 81.42; H, 7.62; N, 10.96. Found: C, 81.21; H,
7.70; N, 10.90.
Acknowledgment. This research has been financially sup-
ported (W. D. O.) by the Council for Chemical Sciences of The
Netherlands Organization for Scientific Research (CW-NWO).
J. Piris and Dr. B. R. Wegewijs (IRI) are gratefully acknowl-
edged for their experimental contribution.
Supporting Information Available: The synthetic proce-
dures for the preparation of 2e and D-B-A compounds D-
[3π3π3π3]A, D[7π3]A, D[11]A, and D[3π7]A. Deconvoluted
fluorescence spectra of D[3π3]A in cycohexane and of D[7π3]-
A in diethyl ether and di-n-butyl ether (PDF). This information
is available free of charge via the Internet at http://pubs.acs.org.

3624 J. Phys. Chem. A, Vol. 107, No. 19, 2003
Oosterbaan et al.
(38) Pasman, P.; Rob, F.; Verhoeven, J. W. J. Am. Chem. Soc. 1982,
104, 5127-5133.
References and Notes
(1) Wasielewski, M. R. Chem. ReV. 1992, 92, 435-461.
(2) Oevering, H.; Paddon-Row: M. N.; Heppener, M.; Oliver, A. M.;
Cotsaris, E.; Verhoeven, J. W.; Hush, N. S. J. Am. Chem. Soc. 1987, 109,
3258-3269.
(3) Bolton, J. R., Mataga, N.,McLendon, G., Eds. Electron Transfer
in Inorganic, Organic, and Biological Systems; Advances in Chemistry
Series 228; American Chemical Society: Washington, DC, 1991.
(4) Barbara, P. F.; Meyer, T. J.; Ratner, M. A. J. Phys. Chem. 1996,
100, 13148-13168.
(5) McConnell, H. M. J. Chem. Phys. 1961, 35, 508-515.
(6) Friesner, R. A.; Won, Y. Biochim. Biophys. Acta 1989, 977, 99-
122.
(7) Tsue, H.; Imahori, H.; Kaneda, T.; Tanaka, Y.; Okada, T.; Tamaki,
K.; Sakata, Y. J. Am. Chem. Soc. 2000, 122, 2279-2288.
(8) Miller, S. E.; Lukas, A. S.; Marsh, E.; Bushard, P.; Wasielewski,
M. R. J. Am. Chem. Soc. 2000, 122, 7802-7810.
(9) Bixon, M.; Jortner, J. J. Chem. Phys. 1997, 107, 5154-5170.
(10) Sumi, H.; Kakitani, T. Chem. Phys. Lett. 1996, 252, 85-93.
(11) Hoogesteger, F. J.; Havenith, R. W. A.; Zwikker, J. W.; Jenneskens,
L. W.; Kooijman, H.; Veldman, N.; Spek, A. L. J. Org. Chem. 1995, 60,
4375-4384.
(12) Marsman, A. W.; Havenith, R. W. A.; Bethke, S.; Jenneskens, L.
W.; Gleiter, R.; van Lenthe, J. H.; Lutz, M.; Spek, A. L. J. Org. Chem.
2000, 65, 4584-4592.
(13) Marsman, A. W.; Havenith, R. W. A.; Bethke, S.; Jenneskens, L.
W.; Gleiter, R.; van Lenthe, J. H. Eur. J. Org. Chem. 2000, 2629-2641.
(14) Bakkers, E. P. A. M.; Roest, A. L.; Marsman, A. W.; Jenneskens,
L. W.; de Jong-van Steensel, L. I.; Kelly, J. J.; Vanmaekelbergh, D. J.
Phys. Chem. B 2000, 104, 7266-7272.
(15) Bakkers, E. P. A. M.; Marsman, A. W.; Jenneskens, L. W.;
Vanmaekelbergh, D. Angew. Chem., Int. Ed. Engl. 2000, 39, 2297-2299.
(16) Holmlin, R. E.; Ismagilov, R. F.; Haag, R.; Mujica, V.; Ratner, M.
A.; Rampi, M. A.; Whitesides, G. M. Angew. Chem., Int. Ed. Engl. 2001,
40, 2316-2320.
(17) Davis, W. B.; Svec, W. A.; Ratner, M. A.; Wasielewski, M. R.
Nature 1998, 396, 60-63.
(18) Hoogesteger, F. J.; van Walree, C. A.; Jenneskens, L. W.; Roest,
M. R.; Verhoeven, J. W.; Schuddeboom, W.; Piet, J. J.; Warman, J. M.
Chem. Eur. J. 2000, 6, 2948-2959.
(19) For a related example see: Willemse, R. J. Photoinduced Electron
Transfer in Donor-Acceptor Systems with Redox-Active Bridges. Thesis,
University of Amsterdam, Amsterdam, The Netherlands, 1997.
(20) Wegewijs, B.; Hermant, R. M.; Verhoeven, J. W.; Kunst, A. G.
M.; Rettschnick, R. P. H. Chem. Phys. Lett. 1987, 140, 587-590.
(21) Lauteslager, X. Y.; van Stokkum, I. H. M.; van Ramesdonk, H. J.;
Brouwer, A. M.; Verhoeven, J. W. J. Phys. Chem. A 1999, 103, 653-659.
(22) Wegewijs, B.; Verhoeven, J. W. Long-Range Charge Separation
in Solvent-Free Donor-Bridge-Acceptor Systems. In Electron-Transfer from
Isolated Molecules to Biomolecules, Part One; Jortner, J.; Bixon, M., Eds.;
Advances in Chemical Physics 106; John Wiley & Sons: New York, 1999.
(23) Schuddeboom, W.; Scherer, T.; Warman, J. M.; Verhoeven, J. W.
J. Phys. Chem. 1993, 97, 13092-13098.
(39) Hermant, R. M.; Bakker, N. A. C.; Scherer, T.; Krijnen, B.;
Verhoeven, J. W. J. Am. Chem. Soc. 1990, 112, 1214-1221.
(40) Scherer, T.; Hielkema, W.; Krijnen, B.; Hermant, R. M.; Eijckelhoff,
C.; Kerkhof, F.; Ng, A. K. F.; Verleg, R.; van der Tol, E. B.; Brouwer, A.
M.; Verhoeven, J. W. Recl. TraV. Chim. Pays-Bas 1993, 112, 535-548.
(41) Excitation spectra suggest that the remaining “local emission” is,
at least in part, due to emission from molecules which lack the acceptor
and are present as a minor impurity.
(42) Marcus, R. A. J. Phys. Chem. 1989, 93, 3078-3086.
(43) It should be noted that for the fitting procedure the experimental
spectra I(λ) were converted to a cm-1 scale without correction for the
intensity (which is usually done via I(ν) = λ2I(λ), see Lakowicz, J. R.
Principles of Fluorescence Spectroscopy; Plenum Press: New York, 1983;
Chapter 2) since this correction will shift the maxima. Without correction
the fitted maxima νmax in cm-1 are directly comparable to experimental
maxima νf.
(44) Fraser, R. D. B.; Suzuki, E. Anal. Chem. 1969, 41, 37-39.
(45) Variation of the band widths ∆ν in the range (4.75-5.50) × 103
cm-1 did not result in shifts of maxima νmax larger than ca. 300 cm-1 relative
to the values in Table 4.
(46) The contribution of a small amount of CCT emission in these
solvents cannot be excluded.
(47) Beens, H.; Knibbe, H.; Weller, A. J. Chem. Phys. 1967, 47, 1183-
1184.
(48) Lippert, E. Z. Elektrochem. 1957, 61, 962-975.
(49) Here it is assumed that F3 varies as the product of Rc and the
2
2
molecule’s cross section and that µe varies as the product of Rc and e2.
(50) Shida, T. Electronic Absorption Spectra of Radical Ions; Physical
Sciences Data 34; Elsevier: Amsterdam, 1988.
(51) Roest, M. R.; Lawson, J. M.; Paddon-Row: M. N.; Verhoeven, J.
W. Chem. Phys. Lett. 1994, 230, 536-542.
(52) The fluorescence maxima in THF were omitted in these fits. Their
position and the polar nature of the solvent are in line with the occurrence
of ECT fluorescence.
(53) A comparison with D[3π3]A instead of D[3]A yields distances of
2.1-2.3 Å.
(54) One of the referees remarked that the energy of a virtual state is
not solvent dependent since there is no driving force for solvent alignment.
However, in the (relaxed) CCT state, solvent reorganization has already
occurred around the acceptor and this will also have a stabilizing effect on
the virtual [3π•+3]A•- state. Thus, the [3π•+3]A•- level is also solvent
dependent.
(55) van Walree, C. A.; Roest, M. R.; Schuddeboom, W.; Jenneskens,
L. W.; Verhoeven, J. W.; Warman, J. M.; Kooijman, H.; Spek, A. L. J.
Am. Chem. Soc. 1996, 118, 8395-8407.
(56) van Dijk, S. I.; Wiering, P. G.; van Staveren, R.; van Ramesdonk,
H. J.; Brouwer, A. M.; Verhoeven, J. W. Chem. Phys. Lett. 1993, 214,
502-506.
(57) Wasielewski, M. R.; Niemczyk, M. P.; Johnson, D. G.; Svec, W.
A.; Minsek, D. W. Tetrahedron 1989, 45, 4785-4806.
(58) de Haas, M. P.; Warman, J. M. Chem. Phys. 1982, 73, 35-53.
(59) Schuddeboom, W. Photophysical Properties of Opto-Electric
Molecules Studied by Time-Resolved Microwave Conductivity. Thesis,
Delft University of Technology, Delft, The Netherlands, 1994.
(60) Weller, A. Z. Phys. Chem., Neue Folge 1982, 133, 93-98.
(61) Rehm, D.; Weller, A. Ber. Bunsen-Ges. Phys. Chem. 1969, 73,
834-839.
(62) For the absorption spectrum, this is a plot of ꢀ(ν)/ν vs. ν. For
fluorescence spectra the wavenumber dependency of the Einstein coefficient
of spontaneous emission has to be taken into account and the corrected
spectrum is represented by I(ν)/ν3 vs ν, where I(ν) = I(λ)λ2.
(63) Bolton, J. R.; Archer, M. D. Basic Electron-Transfer Theory. In
AdVances in Chemistry Series 228: Electron Transfer in Inorganic, Organic,
and Biological Systems; Bolton, J. R., Mataga, N., McLendon, G., Eds.;
American Chemical Society: Washington, DC, 1991.
(24) Hoogesteger, F. J. Oligo(Cyclohexylidenes), Development of Novel
Functional and Organized Materials. Thesis, Utrecht University, Utrecht,
The Netherlands, 1996.
(25) Sa´nchez, I. H.; Ortega, A.; Garc´ıa, G.; Larraza, M. I.; Flores, H. J.
Synth. Commun. 1985, 15, 141-149.
(26) Wells, A. P.; Kitching, W. J. Chem. Soc., Perkin Trans. 1 1995,
527-535.
(27) Trost, B. M.; Tamaru, Y. J. Am. Chem. Soc. 1977, 99, 3101-3113.
(28) Paddon-Row, M. N. Acc. Chem. Res. 1994, 27, 18-25.
(29) Shephard, M. J.; Paddon-Row, M. N. J. Phys. Chem. A 2000, 104,
11628-11635.
(30) Paulson, B.; Pramod, K.; Eaton, P.; Closs, G.; Miller, J. R. J. Phys.
Chem. 1993, 97, 13042-13045.
(31) Pasman, P.; Koper, N. W.; Verhoeven, J. W. Recl. TraV. Chim.
Pays-Bas 1982, 101, 363-364.
(64) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265-
322.
(32) Closs, G. L.; Calcaterra, L. T.; Green, N. J.; Penfield, K. W.; Miller,
J. R. J. Phys. Chem. 1986, 90, 3673-3683.
(65) de Rege, P. J. F.; Williams, S. A.; Therien, M. J. Science 1995,
269, 1409-1413.
(33) Asveld, E. W. H.; Kellogg, R. M. J. Org. Chem. 1982, 47, 1250-
(66) Kroon, J.; Verhoeven, J. W.; Paddon-Row: M. N.; Oliver, A.
Angew. Chem., Int. Ed. Engl. 1991, 30, 1358-1361.
1257.
(34) Oosterbaan, W. D.; Koeberg, M.; Piris, J.; Havenith, R. W. A.;
van Walree, C. A.; Wegewijs, B. R.; Jenneskens, L. W.; Verhoeven, J. W.
J. Phys. Chem. A 2001, 105, 5984-5989.
(67) Brandsma, L.; Verkruijsse, H. D. PreparatiVe Polar Organometallic
Chemistry 1; Springer-Verlag: Berlin, 1987; pp 7-8.
(68) Eaton, D. F. Pure Appl. Chem. 1988, 60, 1107-1114.
(69) Willemse, R. J.; Piet, J. J.; Warman, J. M.; Hartl, F.; Verhoeven,
J. W.; Brouwer, A. M. J. Am. Chem. Soc. 2000, 122, 3721-3730.
(70) Marsman, A. W. Functionalized Oligo(Cyclohexylidenes). Semi-
rigid Molecular Rods for Crystal Engineering, Through-Bond Orbital
Interactions and Self-Assembly. Thesis, Utrecht University, Utrecht, The
Netherlands, 1999, Chapter 8.
(35) Oosterbaan, W. D.; Havenith, R. W. A.; van Walree, C. A.;
Jenneskens, L. W.; Gleiter, R.; Kooijman, H.; Spek, A. L. J. Chem. Soc.,
Perkin Trans. 2 2001, 1066-1074.
(36) Hoogesteger, F. J.; van Lenthe, J. H.; Jenneskens, L. W. Chem.
Phys. Lett. 1996, 259, 178-184.
(37) Havenith, R. W. A.; van Lenthe, J. H.; Jenneskens, L. W.;
Hoogesteger, F. J. Chem. Phys. 1997, 225, 139-152.